History of Reversi: Waterman, Othello and the 2023 Proof

Two claims to one idea

Lewis Waterman put Reversi on sale in May 1883. That date comes from his own sworn statement in a later court case, where he said he had worked out the rules during 1882 and early 1883. The firm Jaques & Son published the game. It needed an ordinary chessboard and sixty-four counters, each dark on one side and light on the other. A player who placed a counter turned over every enemy counter caught in a line between it and another of their own.

A rival maker, F. H. Ayres of London, told a different story. Its 1889 handbook says the game was first introduced in 1870 as The Game of Annexation by John W. Mollett, on a board shaped like a cross. According to the handbook, the game later moved to an ordinary draughtboard as Annex, a Game of Reverses, and only then became known as Reversi. Waterman’s authorized handbook answered that Annexation was an imitation that had been “re-christened” with his game’s name.

Historians can’t settle the question from what survives. The 1870 date rests on Ayres’s own handbook, printed the year after the two sides met in court, and in court Ayres said only that he had sold Annexation since 1876. Even the inventor’s name is uncertain. The handbook’s text credits John W. Mollett, and the box label printed in the same book reads T. W. Mollett. What is clear is that Waterman’s board and Waterman’s name are the ones that lasted.

Lyra, in a violet knit polo, holds up a black disc beside a green Reversi board with seven discs of each color. — Lyra hosts Reversi on LarkSpark. The green board came later, with Othello. Waterman’s Reversi was played on an ordinary chessboard.

Reversi in print and in court

On August 21, 1886, the London weekly The Saturday Review printed an article titled Reverses. It is often cited as the first dependable description of the game in print. The writer called the game “not at present widely known” and explained all of it: a draughtboard, sixty-four two-colored counters, and the rule that you must turn every counter you enclose. The advice still holds. Get a corner if you can, the writer said, and shun the three squares next to it. Today those are the x-square and the two C-squares.

More followed. Waterman’s authorized handbook says a series of articles ran in The Queen newspaper in the spring of 1888, and that a handbook approved by the inventor came out that autumn. By 1890 its author, who wrote as Berkeley, could say that few games had taken such a hold on the public in so short a time.

Waterman registered the word Reversi as a trademark on October 18, 1887. In March 1888 he sued Ayres, who was selling Annex with the words “a game of reverses” on the box. Mr. Justice Kay sided with Waterman on April 20. The Court of Appeal overturned that decision on May 16, 1888. The judges held that Reversi plainly suggested reversing, so the word described the game and could not be registered as a “fancy word.” The mark came off the register, and any maker could sell a game called Reversi.

An empty board and thirty-two counters each

The 1890 handbook Reversi and Go Bang, published in New York with Waterman’s approval, sets out the game as Victorians played it. The board started empty. Black placed a counter on any of the four central squares, White took one of the three that were left, and the two carried on until all four were filled. That allowed two different starts. The colors could sit on diagonals, as they do today, or side by side.

Each player also owned exactly thirty-two counters. A player with no move had the turn postponed, much like a modern pass. The trouble came late in the game. If you had missed turns, your opponent ran out of counters first, and you could then play several moves in a row with the ones still in your hand. The game ended when neither player could move, either because no legal square was left or because the counters were gone.

The game traveled, though the record is thin. The World Othello Federation’s history lists production by the German publisher Ravensburger in 1893 and says Reversi reached Japan by 1907 under the name Genpei Go. An article on the same site by former world champion Ben Seeley doubts that Reversi was ever played in Japan. Both accounts agree on what happened next. The fad faded, no clubs or tournaments formed, and for decades Reversi was treated as a children’s game.

Goro Hasegawa and a new name

In November 1976, Time magazine introduced its readers to Goro Hasegawa, a forty-four-year-old Japanese salesman, and said he had created the game Othello in 1971. He took it to a manufacturer in 1972. The toy company Tsukuda Original began making and selling Othello in 1973, and the first All Japan Championship was played that year. Hasegawa’s father, a scholar of English literature, suggested the name from Shakespeare’s play for the game’s dramatic reversals.

The set looked new. Othello came with a green board of sixty-four squares and discs that were black on one side and white on the other. Hasegawa later wrote of “the beautiful contrast of the white and the black on a green board.” He went on to lead the Japan Othello Association, which promoted the game and ran its tournaments. According to an article by world champion Takeshi Murakami, Tsukuda had sold about 21.5 million sets by 2000.

How much Hasegawa invented is disputed. He said the idea came to him as a schoolboy in postwar Japan, playing with milk-bottle caps, and his obituaries credit him as the game’s creator. Murakami’s account grants that similar games existed in the nineteenth century and credits Hasegawa with the name, the design and the rules that prevailed. The Solitaire Laboratory’s history is blunter and says Othello is identical to Reversi as its author learned it. The sources cited here don’t show whether Hasegawa knew the older game first.

What Othello changed

Set the 1890 handbook beside a modern rulebook and three differences stand out. Every game now begins from the same four discs, with each color on a diagonal. Black always moves first. And the sixty-four discs form one shared supply, so nobody can run out while there are squares left to play.

Passing didn’t change. Victorian players already skipped a turn when they had no move. What changed is what a pass leads to. With a shared supply, play simply continues until neither side can move, and the last move in each region goes to whoever earns it. That is the basis of parity, an idea the old thirty-two-counter limit would have scrambled.

The fixed start matters most. Seeley points out that under the old rules either player could force the side-by-side setup, which plays differently from the diagonal one. With one starting position, an opening can be named, studied and compared from the first move. These are the rules LarkSpark uses. The site calls the game Reversi because Othello is a trademark, and the 1888 ruling left the older name free for anyone.

A minute to learn

In 1975 Gabriel Industries, then a subsidiary of CBS, picked up the right to sell Othello in the United States. By November 1976 Time was describing a craze. A test lot of 3,000 sets priced at $9 had sold out at Woodward & Lothrop in Washington, and Othello was the top seller in the game department at F.A.O. Schwarz in Manhattan. The magazine put the number of players in Japan at about 25 million.

The game arrived with a champion and a slogan. Fumio Fujita, a twenty-seven-year-old barber from outside Tokyo, toured to promote it. In England he beat the chess grandmaster Tony Miles two games to one on closed-circuit television. In Washington he lost to a government lawyer named Mark Weinberg. The game’s motto promised a minute to learn and a lifetime to master. Murakami’s article says about 15 million sets had been sold outside Japan by 2000.

Five players in Tokyo

The first World Othello Championship was played in Tokyo on October 28 and 29, 1977. Five players took part, from Japan, Norway, the United States, the United Kingdom and the Philippines. Hiroshi Inoue of Japan won all four of his round-robin games and then beat Thomas Heiberg of Norway 34–30 in a one-game final. The championship has been held every year since, apart from 2020 and 2021, when the pandemic cancelled it.

For its first decades the event was run by a committee of the companies behind the game. The World Othello Federation was founded in Reykjavik in 2005 by twenty-one national federations and three companies, and it took full responsibility for the championship in 2012. Japanese players have won far more titles than anyone else. Hideshi Tamenori holds the record with seven, the first in 1986 and the last in 2006.

From The Moor to Logistello

In June 1980, Northwestern University hosted the first tournament to set Othello programs against people. Eight competitors were invited: six programs, world champion Hiroshi Inoue and U.S. champion Jonathan Cerf. Inoue won the event, but he lost one game to The Moor, a program from David Levy’s company in London. Cerf lost a game to a program as well.

Paul Rosenbloom’s IAGO, written at Carnegie Mellon University, finished fifth at Northwestern. Seven months later it won the Santa Cruz Open, a tournament for programs only, with eight wins from eight games. Rosenbloom’s paper on IAGO measures things this guide teaches, including mobility and stable discs. BILL, written at the same university by Kai-Fu Lee and Sanjoy Mahajan, first appeared in November 1985 and won its tournament without losing a game.

The decisive match came on August 4 to 7, 1997, at the NEC Research Institute in Princeton, New Jersey. Logistello, written by Michael Buro, played six games against the reigning world champion, Takeshi Murakami, and won all six. Since then the strongest programs have been tools for study. Zebra, a free program begun in June 1997, carries an opening book of more than 500,000 positions. Players use programs like it to check an opening or to count out an endgame.

What it means to solve the game

On June 8, 1993, Joel Feinstein of Nottingham University posted a result to the newsgroup rec.games.abstract. He had worked out Othello on a 6×6 board. With perfect play the second player wins, 16–20 from the first player’s side. The main line took about two weeks of computer time, and the whole search examined tens of billions of positions. The tiny 4×4 board is a win for the second player too.

The full board took thirty more years. On October 30, 2023, Hiroki Takizawa of Preferred Networks posted a paper titled Othello is Solved. It reports a computer proof that the 8×8 game is a draw when neither side makes a mistake. The paper estimates that the game has about 10²⁸ possible positions. The proof didn’t visit them all. It used a modified version of Richard Delorme’s program Edax to solve about 1.5 billion positions with thirty-six empty squares, enough to fix the result from the start.

Takizawa calls this a weak solution. The value of the starting position is known, along with a strategy that guarantees each side at least a draw from there. It doesn’t give the best move in every position that could arise. The paper is a preprint that rests on a very large computation, and its author has published the software so that others can check it. None of this changes the game across a board. Nobody can memorize the drawing lines, and one loose move still decides a game.

The History of Reversi